View Issue Details

IDProjectCategoryView StatusLast Update
0000252bareos-coreGeneralpublic2015-03-31 14:41
Reporterpmarques Assigned Topstorz  
Status closedResolutionfixed 
Product Version13.2.0 
Summary0000252: Ability to convert a failed partial backup into a valid backup.
DescriptionIt has happened several times to me that after waiting for several hours a backup fails because communications failed or the client crashes. The end result of this is a wasted time and tape/disk space.
It would be interesting to be able to decide (at the job level?) to do commits to the database based on time or in data transferred.
For example a big FULL backup over a slow line can take days, and never complete and abort at 80%, this way the rest can go on the next incremental easily.
Steps To ReproduceStart a big backup. Kill the client when almost finished. You end up with a failed backup and a lot of wasted space in the SD storage.
TagsNo tags attached.




2013-11-19 04:50

reporter   ~0000731

Last edited: 2013-11-19 04:51

When spooling is enabled, the spool file maximum size before it is flushed to the disk/tape can be set. Can a "commit point" easily introduced at that flush time?



2013-12-03 11:43

administrator   ~0000739

Hello pmarques,

you can already restore from a partial backup by selecting the jobid of the failed job:

restore jobid=<jobid>

Of course that can only restore data that was already backed up in the failed job.

To set a failed job to have terminated successful, you could update the job table column jobstatus form 'E' to 'T'.

If you do so, you will still not backup the missing data on the next incremental backup, because the determinatino of what has to be backed up is based on the timestamp.

The only possibility I see to backup the missing files would be to use accurate backup. You could try to use accurate and see if the next incremental backup really gets all data that was not backed up during the first full backup.

I hope this info helps.

best regards,



2013-12-03 16:54

reporter   ~0000740


Thank you for your answer.
The problem for me is not as much as to be able to recover the partial data than to be able to have the partial backup consolidated on the database as is, enabling the normal selection and recovery of individual files from that failed backup instance.

Also, if you have spooling enable, as I do, if the backup fails for whatever reason the partial backup will be deleted for good and it will never make it to the database or to the tapes.




2015-03-31 14:41

administrator   ~0001669

We are not planning to implement that until we find some sponsor for it.

Issue History

Date Modified Username Field Change
2013-11-18 06:40 pmarques New Issue
2013-11-18 11:27 mvwieringen Priority high => low
2013-11-18 11:27 mvwieringen Severity major => feature
2013-11-19 04:50 pmarques Note Added: 0000731
2013-11-19 04:51 pmarques Note Edited: 0000731
2013-11-29 15:58 maik Assigned To => pstorz
2013-11-29 15:58 maik Status new => assigned
2013-12-03 11:43 pstorz Note Added: 0000739
2013-12-03 16:54 pmarques Note Added: 0000740
2015-03-31 14:41 pstorz Note Added: 0001669
2015-03-31 14:41 pstorz Status assigned => closed
2015-03-31 14:41 pstorz Resolution open => fixed